Conlang Random Thread

[bradrn]

The problem with calling it a 'nominative' case is then that the typical object is 'nominative', which goes against the very idea that the nominative is a subject case.

Yes, and yet it’s still common terminology. Whereas I don’t believe I’ve seen ‘direct’ used previously for this sort of situation.

It's more like

ERG: -s
DIR: -0
ACC -m

It is practically and functionally identical, but I think there is subtle analytical nuance.

You mean by that tripartite alignment, which is very rare and is not indicated by the term 'direct' case.

I don’t think you and Ahzoh are talking about anything genuinely different. It’s just a terminological quibble. (And a particularly useless one at that — we all agree on how the case functions, so why are we arguing so much over what to call it?)

[Ahzoh]

To begin with, from the Vrkhazhian's perspective, cases are divided into two categories: syntactic cases and adnominal cases.

You mean from the perspective of your con-people? Because "syntactic cases" doesn't make sense linguistically. And "adnominal case" just means "noun case".

Yes, it says "from the Vrkhazhian's perspective", so it's an emic analysis not an etic one. However, the concept of "syntactic case" is also a (cross)linguistic concept, they are literally cases that mark the syntactic elements/roles of a clause. Nominative, Accusative, and Dative mark the role of a noun in the syntax of the language. Also "syntactic case" as a concept is often contrasted with "lexical cases" like Locatives and Ablatives which often serve as adverbial phrases.

Also, no, adnominal doesn't mean "noun case", it literally means "attached to or modifying a noun"
https://en.wiktionary.org/wiki/adnominal

Cases like the Nominative (and Vocative, which I consider a subcase of Nominative), Accusative, Ergative, and Instrumental cases would be syntactic cases as they can exist independently

Again, this reads like word salad. What does "exist independently" mean?

Adnominal cases would include the Genitive, Equative, Ablative, Locative, Comitative, Ornative (possessing X), and Privative (lacking X) cases, because they do not exist independently and instead serve as modifiers of other nouns and verbs.

Cases that "serve as modifiers of (...) nouns and verbs" - again, word salad.

Independently in that they stand alone and do not modify/qualify/attach to other nouns. It's not word salad, it's just very bad reading comprehension of my previous paragraph.

and contribute to the structure of the sentence.

In what way? Syntactic structure had nothing to do with semantic structure, which seems what you're after here.

I'm referring to syntactic structure to begin with, not semantic structure.

Could say they're adjectivized/adverbialized nouns.

How can a case be an "adjectivized" noun? Or, reading between the lines, do you mean that a noun with a genitive a kind of adjective, like "John's book" is the same type of sentence as "the blue book" (where noth "John's" and "blue" modify "book")? If that's the case (no pun intended), how do these other cases you mention operate? How is "The book is table-LOC", the same type of construction? Or do you have something like "the on-the-table book is blue" in mind? If so, I still think you have a class of adjectivizers instead of noun cases.

Well, "the book COP table-LOC" would not be an example of table-LOC being used adnominally, unless one can analyze table-LOC as behaving adverbially and qualifying the copula.

But, with respect to the conlang, there is no distinction between "woman-NOM at home-GEN" and "woman-NOM home-LOC" or "men-NOM from Rome-GEN" and "men-NOM Rome-ABL" and no distinction between those and something like "woman-NOM tall-NOM". They are all modifying/qualifying a noun and do not denote a syntactic role with respect to the verb.

to include the Instrumental (main purpose is to indicate the theme/secondary object of a ditransitive verb),

That would simply not be called an "instrumentive", unless you want to be purposely contrarian. The instrumental specifies the instrument (what's in a name), and in no semantic analysis is the typical indirect object an instrument.

No, absolutely false. Greenlandic is a secundative alignment language and the theme of a ditransitive is marked with the instrumental case, the same case it uses to mark instruments. It does not have an instrumentive case. Think "he supplied the caravan with goods" where in a secundative language like Greenlandic "goods" would be marked with an instrumental case just like how in an indirective alignment language "goods" would be marked accusative while "caravan" would be marked dative.
https://en.wikipedia.org/wiki/Secundati ... uage#Usage

Anyways, I also have an applicative voice, which promotes oblique arguments to core arguments. So it always feels weird to mark the promoted oblique with accusative while the original core object is seemingly "demoted" from accusative marked to instrumental marked.

In my understanding, when an applicative voice is used to promote oblique arguments to core arguments, the are marked as would be fit for the position they're promoted too. The "original" argument is then moved out of the core, or just omitted, and typically marked with a preposition or the like. It shouldn't (as in, what is common) marked with another core argument noun case.

Absolutely false, the applicative voice is commonly and often (if not always) a valency increasing operation, that means it adds core arguments. See, for example, Swahili. And this is especially the case in my conlang that applicatives always add core arguments, not take them away.

the indirect object (marked or not with a dative) is a core argument, not an oblique. So an oblique isn't marked with the dative, but rather with an oblique case or a preposition.

I'm clearly talking about the promoted oblique, obviously it's not an oblique anymore, but I'm referring to it in the sense of "the object-formerly-an-oblique-argument-now-a-core-argument".

Applicatives decrease valency, I don't see what an "applicative voice-marked ditransitive" would be. (And causatives are a whole other ballpark.)

Again, absolutely false. Applicatives do not and never decrease valency. They either preserve the number or arguments (by adding one and demoting another) or, and is much more common, they simply add core arguments. Again, see Swahili.

There's literally a feature map thing on WALS about whether a given language adds applicative markers to intransitive bases only, transitive bases only, or both. Why the ever loving fuck would there ever be such a language that marks applicatives on intransitive bases only if an applicative is a valency decreasing operation. It's clearly a valency increasing (or preserving) operation, like causatives. And there's even papers talking about Swahili's syncretism between the causative marker and the instrumental applicative marker.

This whole comment of yours should be embarrassing with the levels of poor reading comprehension and confidently incorrect going on.

[Ahzoh]

I agree that 'instrumental' here is probably not the best of name per se.

It's pretty much the official designation for "case that marks the secondary object of a secundative-aligned language". For example, Greenlandic uses the instrumental case, that it also uses to mark instrumental obliques, to mark the secondary object/direct object. And even in English you see something analogous with a verb like "provide X with Y" where "with" would be akin to an instrumental case.

The instrumental is the secundative counterpart to the indirect alignment's dative.

[Travis B.]

The problem with calling it a 'nominative' case is then that the typical object is 'nominative', which goes against the very idea that the nominative is a subject case.

Yes, and yet it’s still common terminology. Whereas I don’t believe I’ve seen ‘direct’ used previously for this sort of situation.

Oh I believe people use 'nominative' that way -- that does not mean it is good terminology.

As for 'direct' it just is a neutral term for 'least marked case' while leaving open that it could be used for more than one of intransitive subject, agent, or patient.

It's more like

ERG: -s
DIR: -0
ACC -m

It is practically and functionally identical, but I think there is subtle analytical nuance.

You mean by that tripartite alignment, which is very rare and is not indicated by the term 'direct' case.

I don’t think you and Ahzoh are talking about anything genuinely different. It’s just a terminological quibble. (And a particularly useless one at that — we all agree on how the case functions, so why are we arguing so much over what to call it?)

Personally I just don't like calling it 'nominative' and think that 'nominative-absolutive' (eight syllables!) is longwinded.

[Travis B.]

Okay, this is what the Wiki has to say about the 'direct' case:

A direct case (abbreviated DIR) is a grammatical case used with all three core relations: both the agent and patient of transitive verbs and the argument of intransitive verbs, though not always at the same time. The direct case contrasts with other cases in the language, typically oblique or genitive.

[bradrn]

The problem with calling it a 'nominative' case is then that the typical object is 'nominative', which goes against the very idea that the nominative is a subject case.

Yes, and yet it’s still common terminology. Whereas I don’t believe I’ve seen ‘direct’ used previously for this sort of situation.

Oh I believe people use 'nominative' that way -- that does not mean it is good terminology.

I actually don’t have a problem with it. There is a certain sense in which ‘nominative’ and ‘absolutive’ should be considered a single case — in both systems, they are almost always the least marked case, and the case used for the intransitive subject. Even more convincingly, often the nominative- or absolutive-marked argument has special syntactic properties, shared with nominative or absolutive arguments in other languages but different to the properties of all other arguments. So I think it’s very reasonable to use ‘nominative’ (or ‘absolutive’!) for a case covering both functions.

Okay, this is what the Wiki has to say about the 'direct' case:

A direct case (abbreviated DIR) is a grammatical case used with all three core relations: both the agent and patient of transitive verbs and the argument of intransitive verbs, though not always at the same time. The direct case contrasts with other cases in the language, typically oblique or genitive.

But the usual implication here is that a direct case does not contrast with any other core argument cases. I’m uncomfortable with the idea of a ‘direct’ case which is not always used for all core arguments.

[Travis B.]

Yes, and yet it’s still common terminology. Whereas I don’t believe I’ve seen ‘direct’ used previously for this sort of situation.

Oh I believe people use 'nominative' that way -- that does not mean it is good terminology.

I actually don’t have a problem with it. There is a certain sense in which ‘nominative’ and ‘absolutive’ should be considered a single case — in both systems, they are almost always the least marked case, and the case used for the intransitive subject. Even more convincingly, often the nominative- or absolutive-marked argument has special syntactic properties, shared with nominative or absolutive arguments in other languages but different to the properties of all other arguments. So I think it’s very reasonable to use ‘nominative’ (or ‘absolutive’!) for a case covering both functions.

'Nominative' implies nominative-accusative alignment while 'absolutive' implies ergative-absolutive alignment, so they are not neutral terms. This is problematic in a language were the least marked case does not neatly line up with S and A (as in a 'nominative' case in nominative-accusative alignment) or with S and O (as in an 'absolutive' case in ergative-absolutive alignment). If the least marked case is the default case for both A and O, then it is neither a 'nominative' nor an 'absolutive' case. Of course, 'nominative-absolutive' neatly fits this, but as mentioned it is not succinct at all.

Okay, this is what the Wiki has to say about the 'direct' case:

A direct case (abbreviated DIR) is a grammatical case used with all three core relations: both the agent and patient of transitive verbs and the argument of intransitive verbs, though not always at the same time. The direct case contrasts with other cases in the language, typically oblique or genitive.

But the usual implication here is that a direct case does not contrast with any other core argument cases. I’m uncomfortable with the idea of a ‘direct’ case which is not always used for all core arguments.

My usage of 'direct' here is that the least marked situation is for S, A, and O to all be a shared unmarked case, including both A and O within the same clause. Of course, A and O are not required to be that case, but that is only so in more marked usages, e.g. inanimate agents and animate objects.

[bradrn]

'Nominative' implies nominative-accusative alignment

I’ll stop you right there — it certainly does not. For one thing, there are the many aforementioned cases where ‘nominative’ implies split-ergativity. That’s not even the only such situation: for instance, there are many fluid-S systems with one nominative case and one absolutive case.

My usage of 'direct' here is that the least marked situation is for S, A, and O to all be a shared unmarked case, including both A and O within the same clause. Of course, A and O are not required to be that case, but that is only so in more marked usages, e.g. inanimate agents and animate objects.

Yes, I understand your usage. I’m just pointing out that the linguistic literature usually uses it to mean something different.

[Travis B.]

'Nominative' implies nominative-accusative alignment

I’ll stop you right there — it certainly does not. For one thing, there are the many aforementioned cases where ‘nominative’ implies split-ergativity. That’s not even the only such situation: for instance, there are many fluid-S systems with one nominative case and one absolutive case.

You're right -- I forgot about split-ergativity and fluid-S.

My usage of 'direct' here is that the least marked situation is for S, A, and O to all be a shared unmarked case, including both A and O within the same clause. Of course, A and O are not required to be that case, but that is only so in more marked usages, e.g. inanimate agents and animate objects.

Yes, I understand your usage. I’m just pointing out that the linguistic literature usually uses it to mean something different.

It seems like the only accurate alternative here is 'nominative-absolutive', but I have already expressed why I don't like that term.

[bradrn]

My usage of 'direct' here is that the least marked situation is for S, A, and O to all be a shared unmarked case, including both A and O within the same clause. Of course, A and O are not required to be that case, but that is only so in more marked usages, e.g. inanimate agents and animate objects.

Yes, I understand your usage. I’m just pointing out that the linguistic literature usually uses it to mean something different.

It seems like the only accurate alternative here is 'nominative-absolutive', but I have already expressed why I don't like that term.

Yes, there sadly aren’t any truly good terms for this. The difficulty is that it’s really a continuum, from nominative-accusative to ergative-absolutive, as well as to Austronesian and inverse alignments, and marked-nominative alignments, and so on. All of these clearly have one privileged argument, but when does that argument stop being ‘nominative’ and start being something else? There’s no clear way to tell, and linguists can never agree on a convention — witness the endless debates on what to call the privileged argument in Austronesian-type alignments.

[Travis B.]

Yes, I understand your usage. I’m just pointing out that the linguistic literature usually uses it to mean something different.

It seems like the only accurate alternative here is 'nominative-absolutive', but I have already expressed why I don't like that term.

Yes, there sadly aren’t any truly good terms for this. The difficulty is that it’s really a continuum, from nominative-accusative to ergative-absolutive, as well as to Austronesian and inverse alignments, and marked-nominative alignments, and so on. All of these clearly have one privileged argument, but when does that argument stop being ‘nominative’ and start being something else? There’s no clear way to tell, and linguists can never agree on a convention — witness the endless debates on what to call the privileged argument in Austronesian-type alignments.

With regard to case-marking this alignment is different in that there really is no privileged argument -- both A and O can be unmarked or marked independent of one another, and S is always unmarked. It is different from plain split-ergativity in that all four combinations of unmarked and marked A and unmarked and marked O are valid instead of just having either an unmarked A and a marked O or a marked A and an unmarked O. Of course, underlyingly there are likely to be signs of privileging one argument, with things like syntactic pivot and so on.

[Travis B.]

Hmm... Morphosyntactic alignment on the Wiki does not even mention this sort of alignment. Of course, it also barely mentions split ergativity, which is much more common than people give it credit for.

[Travis B.]

Of course, Wikipedia is really known for the quality of its linguistics articles... Differential argument marking is somewhat better overall in this regard.

I just got an idea that I have lifted from Hindi ─ yes, from reading the Wiki ─ definiteness can be expressed for inanimate O and adpositional arguments by placing them in the accusative case rather than their default of the, well, nominative-absolutive case. Likewise, by extension, definiteness could be expressed for animate A by placing it in the ergative case.

[Ahzoh]

It seems like the only accurate alternative here is 'nominative-absolutive', but I have already expressed why I don't like that term.

I've seen it just called Nominative, such as on the wiki page for Hittite Grammar.

[bradrn]

With regard to case-marking this alignment is different in that there really is no privileged argument -- both A and O can be unmarked or marked independent of one another, and S is always unmarked. It is different from plain split-ergativity in that all four combinations of unmarked and marked A and unmarked and marked O are valid instead of just having either an unmarked A and a marked O or a marked A and an unmarked O. Of course, underlyingly there are likely to be signs of privileging one argument, with things like syntactic pivot and so on.

No, what you describe is simply split-ergativity with a split based on animacy. (Incidentally I dislike the term ‘split-ergativity‘ because the name makes it sound like ergativity is somehow privileged, which it’s not.) It is indeed a characteristic of most split-ergative systems that ‘all four combinations […] are valid’.

[Travis B.]

With regard to case-marking this alignment is different in that there really is no privileged argument -- both A and O can be unmarked or marked independent of one another, and S is always unmarked. It is different from plain split-ergativity in that all four combinations of unmarked and marked A and unmarked and marked O are valid instead of just having either an unmarked A and a marked O or a marked A and an unmarked O. Of course, underlyingly there are likely to be signs of privileging one argument, with things like syntactic pivot and so on.

No, what you describe is simply split-ergativity with a split based on animacy. (Incidentally I dislike the term ‘split-ergativity‘ because the name makes it sound like ergativity is somehow privileged, which it’s not.) It is indeed a characteristic of most split-ergative systems that ‘all four combinations […] are valid’.

The Wiki article on split ergativity makes it sound that split ergativity involves conditioned choices between nominative-accusative and ergative-absolutive alignment and/or multiple different systems that use both simultaneously (e.g. agreement based on one sort of alignment and case-marking based on another sort of alignment). Of course we all know the quality of Wikipedia linguistics articles... This may be why I didn't call this alignment 'split-ergative', as it does not involve a binary choice between nominative-accusative and ergative-absolutive alignment.

[Travis B.]

It seems like the only accurate alternative here is 'nominative-absolutive', but I have already expressed why I don't like that term.

I've seen it just called Nominative, such as on the wiki page for Hittite Grammar.

Apparently it is common to call this 'nominative' in Indic linguistics and 'absolutive' in Philippine linguistics.

[bradrn]

The Wiki article on split ergativity

Firstly, Wikipedia is rubbish.

split ergativity involves conditioned choices between nominative-accusative and ergative-absolutive alignment

This is correct though.

and/or multiple different systems that use both simultaneously (e.g. agreement based on one sort of alignment and case-marking based on another sort of alignment).

I don’t think I’ve actually seen people call this ‘split ergativity’, but it’s certainly a closely related thing.

This may be why I didn't call this alignment 'split-ergative', as it does not involve a binary choice between nominative-accusative and ergative-absolutive alignment.

But from your description that’s what it sounded like — nominative-accusative for more animate arguments, ergative-absolutive for less animate arguments.

Apparently it is common to call this 'nominative' in Indic linguistics and 'absolutive' in Philippine linguistics.

I don’t recall seeing ‘absolutive’ in that context.

[Travis B.]

I think where we are getting confused is that by 'choice' I meant choice for the whole clause, not a choice for each argument independent of one another. What I was referring to is definitely split-ergativity if split-ergativity is independent for each argument (i.e. A can have a different alignment from O within the same clause).

[bradrn]

I think where we are getting confused is that by 'choice' I meant choice for the whole clause, not a choice for each argument independent of one another. What I was referring to is definitely split-ergativity if split-ergativity is independent for each argument (i.e. A can have a different alignment from O within the same clause).

It depends on the type of split-ergativity, but when the split is based on animacy, it’s certainly independent for each argument.

(Of course for other kinds of splits that might be logically impossible, e.g. when it’s based on TAM.)